I wish to compare some methods of disease classification. I have calculated sensitivity and specificity of each methods against a "gold standard" for diagnosis Y. I need to calculate the relative risk (RR) of disease Y for a binary exposure X, which is assumed to be measured free of error.
I found formulas to derive RR as a function of sensitivity and specificity of EXPOSURE misclassification (e.g. Flegal et al. American Journal Epidemiology 1986) but I can't find formulas to derive RR as a function of sensitivity and specificity of DISEASE misclassification.
Can anybody help?
Kind regards
Rothman and Greenland, Modern Epidemiology 2nd ed p347-352 deals with both exposure and disease misclassification and gives adjustment formulae in terms of sensitivity and specificity for both cohort and case referent studies.
The measures are related but numerators and denominators are not the same. BUT if you were to measure NOT the risks of illness among each exposure group, but the EXPOSURE rates among ill and not-ill groups, the measures would be drawn from the same contingency table as is used for standard sensitivity and specificity measures.
Dear Massimiliano Bugiani ·
Could you please send me this chapter (the book in not available for me)
Rothman and Greenland, Modern Epidemiology chp 19
- Badges
- Science topic
- Similar topics
- Medicine
- Epidemiology